Supported models

This page lists every model you can use on the distil labs platform, and how to choose between them.

Only models in this catalog are supported

The platform only supports the models listed on this page. If your configuration references a model outside this catalog, training will not work. Always pick student_model_name and teacher_model_name values from the tables below.

Students and teachers

Every training run involves two models with different roles:

• Student model: the small model we fine-tune for your task. This is the model you deploy and run inference with, so pick a size that fits your latency, cost, and hardware constraints.
• Teacher model: the large model used to generate and validate synthetic training data. Its knowledge is distilled into the student. You never deploy the teacher.

Supported student models

Model	Value
Llama 3.2 1B Instruct	Llama-3.2-1B-Instruct
Llama 3.2 3B Instruct	Llama-3.2-3B-Instruct
Llama 3.1 8B Instruct	Llama-3.1-8B-Instruct
SmolLM2 135M	SmolLM2-135M-Instruct
SmolLM2 1.7B	SmolLM2-1.7B-Instruct
FunctionGemma 270M	functiongemma-270m-it
Gemma 3 270M	gemma-3-270m-it
Gemma 3 1B	gemma-3-1b-it
Gemma 3 4B	gemma-3-4b-it
Gemma 4 E2B	gemma-4-E2B-it
Gemma 4 E4B	gemma-4-E4B-it
Qwen3 0.6B	Qwen3-0.6B
Qwen3 1.7B	Qwen3-1.7B
Qwen3 4B	Qwen3-4B-Instruct-2507
Qwen3 8B	Qwen3-8B
Qwen3.5 0.8B	Qwen3.5-0.8B
Qwen3.5 2B	Qwen3.5-2B
Qwen3.5 4B	Qwen3.5-4B
Qwen3.5 9B	Qwen3.5-9B
Liquid LFM2 350M	LFM2-350M
Liquid LFM2 1.2B	LFM2-1.2B
Liquid LFM2 2.6B	LFM2-2.6B
Liquid LFM2.5 350M	LFM2.5-350M
Liquid LFM2.5 1.2B Instruct	LFM2.5-1.2B-Instruct

Choosing a student model

• Smaller students (under 1B parameters) give the fastest and cheapest inference and are the easiest to run on-device, but may need more training data to reach your quality bar.
• Larger students (3B and up) usually reach higher quality out of the box at the cost of slower and more expensive inference.
• When in doubt, start with one of the Qwens, e.g. Qwen3-1.7B, check the evaluation metrics, and only move up in size if the results fall short.

Supported teacher models

Model	Value	Tool calling
GPT OSS 120B	openai.gpt-oss-120b	✓
GPT OSS 120B Thinking	openai.gpt-oss-120b-thinking	✓
GPT OSS 20B	openai.gpt-oss-20b	✓
GPT OSS 20B Thinking	openai.gpt-oss-20b-thinking	✓
DeepSeek R1	deepseek.r1	✗
DeepSeek R1 Thinking	deepseek.r1-thinking	✗
DeepSeek V3.1	deepseek.v3.1	✗
DeepSeek V3.2	deepseek.v3.2	✓
Qwen3 235B A22B	Qwen3-235B-A22B-Instruct-2507	✓
Qwen3 480B A35B Coder	Qwen3-480B-A35B-Coder	✗
Qwen2.5 VL 72B	Qwen2.5-VL-72B-Instruct	✗
ZAI GLM 5	zai.glm-5	✓
Moonshot Kimi K2 Thinking	moonshotai.kimi-k2-thinking	✓
Moonshot Kimi K2.5	moonshotai.kimi-k2.5	✓
MiniMax M2 Thinking	minimax.minimax-m2-thinking	✓

The Tool calling column applies to both the tool-calling-closed-book and multi-turn-tool-calling-closed-book tasks.

The default teacher (openai.gpt-oss-120b) is a strong choice for all task types.

Task compatibility

Not every model works with every task type:

• Tool calling tasks (tool-calling-closed-book and multi-turn-tool-calling-closed-book):
  • Student models: only the Qwen3, Qwen3.5, Llama 3-family, LFM2/LFM2.5, FunctionGemma, and Gemma 4 models are supported.
  • Teacher models: only teachers marked with ✓ in the Tool calling column above.
• Reasoning teacher models (the GPT OSS, DeepSeek, GLM, Kimi, and MiniMax families) require a teacher temperature between 0.5 and 0.7. Configurations that set synthgen.teacher_temperature outside this range will raise a validation error.
• GPT OSS 120B Thinking (openai.gpt-oss-120b-thinking) uses a medium reasoning effort setting by default for enhanced chain-of-thought capabilities.

Using these models in your config

Set the models in the base section of your configuration file, using the exact values from the tables above:

base:
  task: question-answering
  student_model_name: Llama-3.2-1B-Instruct
  teacher_model_name: openai.gpt-oss-120b

See the configuration file reference for all available parameters and defaults.